Currently, high definition videos have become an important technical trend in the field of video surveillance, and cameras with resolution of 720p or 1080p are more and more widely applied in the field of video surveillance. With a continuous increase in camera resolution, the range that a single camera is capable of monitoring is increasingly wide, and image details are clearer; at the same time, an intelligent analysis technology for video images is gradually put into practice. With technical development of hardware devices, hardware performance has been capable of meeting requirements for performing intelligent analysis on multiple regions of interest in a same image, which significantly saves costs as compared with manual surveillance.
An existing video surveillance client generally plays video images from multiple cameras at the same time. However, with an increase in the resolution of video images, total resolution of video images from multiple cameras is often beyond the resolution range of a client's monitor. Taking a 22-inch display as an example, the display generally supports maximum resolution of 1920*1080, that is, the display is only allowed to play 1080p images from one feed. If multiple feeds of 1080p images are played on the monitor at the same time, the images have to be zoomed out. In addition to a playback window, there are also multiple auxiliary function panels on an interface of a typical video surveillance client, such as a title bar, a camera list, and a pan-tilt-zoom control panel, which further reduce a display area of video images. Therefore, images that can be played in the playback window are far smaller than original images.
Especially, when an event (for example, an event triggered by intelligent analysis) occurs in a video image, the region of the image where the event occurs is even smaller because the image is downsized during playback, which makes it inconvenient for a user to view. If observation personnel monitor images with naked eyes, it is hard for the observation personnel to notice the change of details, resulting in missing of key information.
Currently, most of the clients provide a function of zooming in a selected region of an image. That is, a region of a video playback image is selected by sliding a mouse, and the selected region is zoomed in, thereby improving image quality of the region of interest to a certain extent. However, digital zooming of the video image causes loss of some pixel information, therefore affecting the image quality and also affecting the effect of user observation on image details. In addition, if the function of zooming in a selected region is used, in one aspect, a manual operation by the user is needed, and the user may have no time to perform any operation when an event occurs suddenly, and therefore miss the event; in another aspect, if an event occurs in different regions of the image, it is impossible to zoom in several regions at the same time. Therefore, user experience is relatively poor.